Impact of surface roughness on L-band emissivity of the sea ice

Abstract:

In March 2014 a joint experiment IRO2/SMOSice was carried out in the Barents Sea. R/V Lance equipped with meteorological instruments, electromagnetic sea ice thickness probe and engine monitoring instruments, was performing a series of tests in different ice conditions in order to validate the ice route optimization (IRO) system, advising on his route through pack ice. In parallel cal/val activities for sea ice thickness product obtained from SMOS (Soil Moisture and Ocean Salinity mission) L-band radiometer were carried out. Apart from helicopter towing the EMbird thickness probe, Polar 5 aircraft was serving the area during the experiment with L-band radiometer EMIRAD2 and Airborne Laser Scanner (ALS) as primary instruments.

Sea ice Thickness algorithm using SMOS brightness temperature developed at University of Hamburg, provides daily maps of thin sea ice (up to 0.5-1 m) in polar regions with resolution of 35-50 km. So far the retrieval method was not taking into account surface roughness, assuming that sea ice is a specular surface. Roughness is a stochastic process that can be characterized by standard deviation of surface height σ and by shape of the autocorrelation function R to estimate it's vertical and horizontal scales respectively. Interactions of electromagnetic radiation with the surface of the medium are dependent on R and σ and they scales with respect to the incident wavelength. During SMOSice the radiometer was observing sea ice surface at two incidence angles 0 and 40 degrees and simultaneously the surface elevation was scanned with ALS with ground resolution of ~ 0.25 m. This configuration allowed us to calculate σ and R from power spectral densities of surface elevation profiles and quantify the effect of surface roughness on the emissivity of the sea ice. First results indicate that Gaussian autocorrelation function is suitable for deformed ice, for other ice types exponential function is the best fit.